OpenMOSS-Team/MOSS-TTS-Realtime
10711k
1{2 "model_type": "moss_tts_realtime",3 "architectures": [4 "MossTTSRealtime"5 ],6 "auto_map": {7 "AutoConfig": "configuration_mossttsrealtime.MossTTSRealtimeConfig",8 "AutoModel": "modeling_mossttsrealtime.MossTTSRealtime",9 "AutoProcessor": "processing_mossttsrealtime.MossTTSRealtimeProcessor"10 },11 "processor_class": "MossTTSRealtimeProcessor",12 "audio_pad_token": 1024,13 "audio_vocab_size": 1027,14 "dtype": "bfloat16",15 "initializer_range": 0.02,16 "language_config": {17 "_name_or_path": ".",18 "architectures": [19 "Qwen3ForCausalLM"20 ],21 "attention_bias": false,22 "attention_dropout": 0.0,23 "bos_token_id": 151643,24 "dtype": "bfloat16",25 "eos_token_id": 151645,26 "head_dim": 128,27 "hidden_act": "silu",28 "hidden_size": 2048,29 "initializer_range": 0.02,30 "intermediate_size": 6144,31 "layer_types": [32 "full_attention",33 "full_attention",34 "full_attention",35 "full_attention",36 "full_attention",37 "full_attention",38 "full_attention",39 "full_attention",40 "full_attention",41 "full_attention",42 "full_attention",43 "full_attention",44 "full_attention",45 "full_attention",46 "full_attention",47 "full_attention",48 "full_attention",49 "full_attention",50 "full_attention",51 "full_attention",52 "full_attention",53 "full_attention",54 "full_attention",55 "full_attention",56 "full_attention",57 "full_attention",58 "full_attention",59 "full_attention"60 ],61 "max_position_embeddings": 40960,62 "max_window_layers": 28,63 "model_type": "qwen3",64 "num_attention_heads": 16,65 "num_hidden_layers": 28,66 "num_key_value_heads": 8,67 "rms_norm_eps": 1e-06,68 "rope_scaling": null,69 "rope_theta": 1000000,70 "sliding_window": null,71 "use_cache": true,72 "use_sliding_window": false,73 "vocab_size": 15193674 },75 "local_config": {76 "attention_bias": false,77 "attention_dropout": 0.0,78 "audio_pad_token": 1024,79 "audio_vocab_size": 1027,80 "head_dim": 128,81 "hidden_act": "silu",82 "hidden_size": 2048,83 "initializer_range": 0.02,84 "intermediate_size": 6144,85 "max_position_embeddings": 33,86 "model_type": "MossTTSRealtimeLocalTransformer",87 "num_attention_heads": 16,88 "num_hidden_layers": 4,89 "num_key_value_heads": 8,90 "rms_norm_eps": 1e-06,91 "rope_theta": 1000000,92 "rvq": 16,93 "tie_word_embeddings": false,94 "use_cache": true95 },96 "reference_audio_pad": 151654,97 "rvq": 16,98 "text_pad": 151655,99 "tie_word_embeddings": false,100 "transformers_version": "4.57.1",101 "vocab_size": 151936102}103 